#########################################
#   Manejo y Análisis de Datos          #
#########################################
import numpy as np      # Para computación numérica y manejo de arrays
import pandas as pd     # Para manipulación y análisis de datos tabulares (DataFrames)

#################################
#   Visualización de Datos      #
#################################
import matplotlib.pyplot as plt # Para crear gráficos y visualizaciones estáticas
import seaborn as sns           # Para visualizaciones estadísticas atractivas

###################################################
#   Utilidades del Sistema y Matemáticas          #
###################################################
import os                       # Para interactuar con el sistema operativo
import math                     # Para funciones matemáticas básicas
from scipy import stats         # Para funciones estadísticas avanzadas

################################################
#   Aprendizaje Automático (Scikit-learn)      #
################################################

# --- Preprocesamiento y Transformación de Datos ---
from sklearn.preprocessing import LabelEncoder, StandardScaler  # Para codificar y estandarizar características
from sklearn.preprocessing import FunctionTransformer         # Para aplicar funciones personalizadas a los datos
from sklearn.preprocessing import PowerTransformer            # Para aplicar transformaciones de potencia a los datos
from sklearn.decomposition import PCA                         # Para reducción de dimensionalidad
from sklearn.pipeline import Pipeline                         # Para encadenar pasos de procesamiento y modelado

# --- Selección y Evaluación de Modelos ---
from sklearn.model_selection import train_test_split, cross_val_score # Para dividir datos y realizar validación cruzada
from sklearn.model_selection import GridSearchCV                      # Para búsqueda de hiperparámetros
from sklearn.metrics import accuracy_score, confusion_matrix      # Para medir la precisión y ver la matriz de confusión
from sklearn.metrics import classification_report                 # Para un reporte detallado de métricas

# --- Modelos de Clasificación ---
from sklearn.linear_model import LogisticRegression     # Modelo de Regresión Logística
from sklearn.neighbors import KNeighborsClassifier      # Modelo K-Vecinos más Cercanos
from sklearn.svm import SVC                             # Máquinas de Vectores de Soporte (SVM)
from sklearn.naive_bayes import GaussianNB              # Clasificador Naive Bayes Gaussiano
from sklearn.tree import DecisionTreeClassifier         # Modelo de Árbol de Decisión
from sklearn.ensemble import RandomForestClassifier     # Modelo de Bosque Aleatorio (Random Forest)
from sklearn.neural_network import MLPClassifier        # Red Neuronal (Perceptrón Multicapa)

#################################################
#   Librerías Personalizadas del Proyecto 📦    #
#################################################
from acoustic_ml.dataset import load_raw_data       # Para cargar los datos crudos del proyecto
from acoustic_ml.features import create_features    # Para crear nuevas características para el modelo
from acoustic_ml.modeling.predict import load_model, predict # Para cargar y usar un modelo ya entrenado

from acoustic_ml.dataset import load_raw_data                     # Importa la función necesaria del módulo

df = load_raw_data()                                              # Ejecuta la función para cargar los datos
print(f"✓ Dataset cargado: {df.shape[0]} filas y {df.shape[1]} columnas") # Verifica y reporta las dimensiones del dataset

✓ Dataset cargado: 400 filas y 51 columnas

def info_df(df):
    """
    Genera un DataFrame de resumen estilizado con información clave de cada columna.
    """
    print(f"Análisis del DataFrame: {df.shape[0]} filas y {df.shape[1]} columnas")
    
    # 1. Crear el DataFrame de resumen
    resumen = pd.DataFrame({
        'Tipo de Dato': df.dtypes,
        'Valores No Nulos': df.count(),
        'Valores Nulos': df.isnull().sum(),
        '% Nulos': round(df.isnull().sum() * 100 / len(df), 2),
        'Valores Únicos': df.nunique(),
        'Cardinalidad (%)': round(df.nunique() * 100 / len(df), 2)
    })

    # 2. Ordenar por el porcentaje de nulos para priorizar
    resumen = resumen.sort_values(by='% Nulos', ascending=False)

    # 3. Aplicar estilo visual
    styled_resumen = (resumen.style
                      .background_gradient(cmap='viridis', subset=['% Nulos'])
                      .background_gradient(cmap='plasma_r', subset=['Cardinalidad (%)'])
                      .format({'% Nulos': '{:.2f}%', 'Cardinalidad (%)': '{:.2f}%'})
                      .bar(subset=['Valores No Nulos'], color='#5fba7d')
                      .set_caption("Resumen Analítico del Dataset")
                     )
    
    return styled_resumen

# ¡Llama a la función con nuestro DataFrame!
info_df(df)

Análisis del DataFrame: 400 filas y 51 columnas

estadisticas_descriptivas = df.describe()
estadisticas_descriptivas

def descripcion_df(df):
    """
    Genera una tabla de estadísticas descriptivas completa y estilizada 
    para las columnas numéricas.
    """
    # Seleccionar solo columnas numéricas
    df_num = df.select_dtypes(include=['number'])
    
    if df_num.empty:
        print("No se encontraron columnas numéricas en el DataFrame.")
        return

    # Calcular estadísticas básicas
    desc = df_num.describe().T
    
    # Añadir estadísticas avanzadas
    desc['skew'] = df_num.skew()
    desc['kurtosis'] = df_num.kurt()
    desc['median'] = df_num.median()
    
    # Reordenar para que la mediana esté junto a la media
    desc = desc[['count', 'mean', 'median', 'std', 'min', '25%', '50%', '75%', 'max', 'skew', 'kurtosis']]
    
    print(f"Estadísticas Descriptivas para {df_num.shape[1]} columnas numéricas:")
    
    # Aplicar estilo para resaltar valores importantes
    styled_desc = (desc.style
                   .background_gradient(cmap='coolwarm', subset=['skew', 'kurtosis'])
                   .format('{:.2f}')
                   .set_caption("Estadísticas Descriptivas Avanzadas")
                  )
    return styled_desc

# ¡Llama a la función con nuestro DataFrame!
descripcion_df(df)

Estadísticas Descriptivas para 50 columnas numéricas:

def visualizar_distribuciones_numericas(df):
    """
    Crea un histograma y un diagrama de caja para cada columna numérica en el DataFrame,
    con títulos mejorados en 3 líneas y una paleta de colores inspirada en la música turca.
    """
    # Seleccionar solo columnas numéricas
    df_num = df.select_dtypes(include=['number'])
    
    if df_num.empty:
        print("No se encontraron columnas numéricas para visualizar.")
        return
        
    print("Visualización de Distribuciones Numéricas:")
    
    # Paleta de colores inspirada en la música turca
    colores_turcos = ["#E9A000", "#1C4E80", "#A52A2A", "#8B4513", "#4A8C80", "#D36E1B", "#663399"]
    
    for i, col in enumerate(df_num.columns):
        # AQUÍ ESTÁ EL CAMBIO: figsize aumentado para más espacio
        fig, axes = plt.subplots(1, 2, figsize=(10.5, 3.75))
        
        color_base = colores_turcos[i % len(colores_turcos)]

        # --- Histograma con línea de densidad (KDE) ---
        sns.histplot(df[col], kde=True, ax=axes[0], bins=30, color=color_base, edgecolor='white')
        
        # TÍTULO MEJORADO EN TRES LÍNEAS
        axes[0].set_title(f'Distribución de la Variable\n"{col}"\n(Histograma y KDE)', fontsize=10)
        
        mean_val = df[col].mean()
        median_val = df[col].median()
        axes[0].axvline(mean_val, color='#C0C0C0', linestyle='--', linewidth=2, label=f'Media: {mean_val:.2f}')
        axes[0].axvline(median_val, color='#FFD700', linestyle='-', linewidth=2, label=f'Mediana: {median_val:.2f}')
        axes[0].legend(fontsize='small')
        axes[0].set_xlabel(col, fontsize=8)
        axes[0].set_ylabel('Frecuencia', fontsize=8)
        
        # --- Diagrama de caja (Box Plot) ---
        sns.boxplot(x=df[col], ax=axes[1], color=color_base, flierprops=dict(markerfacecolor='#B22222', marker='D'))
        
        # TÍTULO MEJORADO EN TRES LÍNEAS
        axes[1].set_title(f'Análisis de Dispersión\n"{col}"\n(Box Plot y Outliers)', fontsize=10)
        axes[1].set_xlabel(col, fontsize=8)
        
        # Aumentamos el padding para dar más aire a los títulos
        plt.tight_layout(pad=2.0)
        plt.show()

# --- Llama a la función para ver los gráficos ---
visualizar_distribuciones_numericas(df)

Visualización de Distribuciones Numéricas:

def analizar_cardinalidad(df):
    """
    Analiza y muestra la cardinalidad (valores únicos) de cada columna 
    de una manera visualmente impactante.
    """
    # Crear un DataFrame con los conteos y porcentajes de valores únicos
    cardinalidad_df = pd.DataFrame({
        'Valores Únicos': df.nunique(),
        'Cardinalidad (%)': round(df.nunique() * 100 / len(df), 2)
    }).sort_values(by='Cardinalidad (%)', ascending=False)

    print(f"Análisis de Cardinalidad para {len(df.columns)} columnas:")

    # Aplicar estilo para un análisis visual rápido
    styled_df = (cardinalidad_df.style
                 .background_gradient(cmap='magma', subset=['Cardinalidad (%)'])
                 .format({'Cardinalidad (%)': '{:.2f}%'})
                 .bar(subset=['Valores Únicos'], color='#5fba7d', align='zero')
                 .set_caption("Unicidad y Cardinalidad de las Columnas")
                )
    return styled_df

# --- Llama a la función para un análisis enfocado en la cardinalidad ---
analizar_cardinalidad(df)

Análisis de Cardinalidad para 51 columnas:

def analizar_datos_faltantes(df):
    """
    Analiza y muestra los datos faltantes de una manera impactante, mostrando
    únicamente las columnas que contienen valores nulos.
    """
    # Calcula el conteo y porcentaje de nulos por columna
    valores_faltantes = df.isnull().sum()
    porcentaje_faltante = round(valores_faltantes * 100 / len(df), 2)
    
    # Crea un DataFrame con los resultados
    resumen_faltantes = pd.DataFrame({
        'Valores Faltantes': valores_faltantes,
        'Porcentaje (%)': porcentaje_faltante
    })
    
    # Filtra para mostrar solo columnas con datos faltantes y ordena de mayor a menor
    resumen_faltantes = resumen_faltantes[resumen_faltantes['Valores Faltantes'] > 0].sort_values(
        by='Porcentaje (%)', ascending=False
    )
    
    # Si el DataFrame resultante está vacío, ¡felicidades!
    if resumen_faltantes.empty:
        print("🎉 ¡Excelente! No se encontraron datos faltantes en el DataFrame.")
        return
        
    print(f"Análisis de Datos Faltantes: {len(resumen_faltantes)} de {len(df.columns)} columnas tienen valores nulos.")
    
    # Aplica estilo para resaltar los problemas
    styled_resumen = (resumen_faltantes.style
                      .background_gradient(cmap='Reds', subset=['Porcentaje (%)'])
                      .format({'Porcentaje (%)': '{:.2f}%'})
                      .bar(subset=['Valores Faltantes'], color='#d65f5f', align='zero')
                      .set_caption("Columnas con Datos Faltantes")
                     )
    
    return styled_resumen

# --- Llama a la función para un diagnóstico preciso de datos faltantes ---
analizar_datos_faltantes(df)

🎉 ¡Excelente! No se encontraron datos faltantes en el DataFrame.

def countplot_sentimientos(df, col='Class'):
    """
    Versión del gráfico brutal con un ajuste en los márgenes del eje X
    y el título más cerca del gráfico.
    """
    # --- 1. Preparación de Datos y Paleta de Colores Suavizada ---
    colores_suaves = ["#C89F9C", "#A2D4AB", "#F0E2B6", "#B4A2D4"]
    data = df[col].value_counts()
    order = data.index
    total = len(df[col])
    palette = colores_suaves[:len(order)]

    # --- 2. Creación de la Figura con Tema Oscuro ---
    fig, ax = plt.subplots(figsize=(12, 7))
    fig.patch.set_facecolor('#1E1E1E')
    ax.set_facecolor('#1E1E1E')

    # --- 3. Dibujo de Barras con Gradiente ---
    for i, (category, count) in enumerate(data.items()):
        base_color = palette[i]
        gradient = np.linspace(0.8, 1.0, 256).reshape(-1, 1)
        rgb_color = np.array([int(base_color.lstrip('#')[j:j+2], 16) / 255.0 for j in (0, 2, 4)])
        gradient_colors = plt.cm.colors.ListedColormap(rgb_color * gradient)
        ax.imshow(np.arange(256).reshape(-1, 1), cmap=gradient_colors, 
                  extent=[i - 0.4, i + 0.4, 0, count], aspect='auto', interpolation='bilinear')

    # --- 4. Anotaciones y Tipografía de Alto Impacto ---
    for i, count in enumerate(data):
        percentage = f'{(count / total) * 100:.1f}%'
        ax.text(i, count - (0.05 * data.max()), percentage, 
                ha='center', va='center', color='white', fontsize=14, weight='bold',
                bbox=dict(facecolor='black', alpha=0.3, boxstyle='round,pad=0.2', edgecolor='none'))

    # --- 5. Limpieza y Estilo Final ---
    ax.set_xticks(range(len(order)))
    ax.set_xticklabels(order, fontsize=12, color='white', weight='medium')
    ax.set_ylim(0, data.max() * 1.1)
    
    ax.set_xlim(-0.5, len(order) - 0.5)
    
    ax.spines['top'].set_visible(False)
    ax.spines['right'].set_visible(False)
    ax.spines['left'].set_visible(False)
    ax.spines['bottom'].set_color('#444444')
    ax.tick_params(axis='x', colors='white', length=0)
    ax.set_yticks([])
    ax.set_ylabel('')
    ax.set_xlabel('')
    
    plt.title('Distribución de Clases', fontsize=20, color='#DDDDDD', weight='bold', pad=1)
    plt.show()

# --- Llama a la función ajustada ---
countplot_sentimientos(df)

def visualizar_violines(df, categoria_col='Class'):
    """
    Crea una cuadrícula de gráficos de violín
    """
    df_num = df.select_dtypes(include=np.number)
    if 'Track ID' in df_num.columns:
        df_num = df_num.drop(columns=['Track ID'])

    if categoria_col not in df.columns:
        print(f"Error: La columna '{categoria_col}' no se encuentra en el DataFrame.")
        return

    # Paleta de colores y ajuste dinámico al número de categorías
    n_categorias = df[categoria_col].nunique()
    colores_suaves = ["#C89F9C", "#A2D4AB", "#F0E2B6", "#B4A2D4"]
    palette_ajustada = colores_suaves[:n_categorias]
    
    n_variables = len(df_num.columns)
    n_cols_grid = 3
    n_rows_grid = (n_variables + n_cols_grid - 1) // n_cols_grid

    # Crear la figura con tema oscuro
    fig, axes = plt.subplots(n_rows_grid, n_cols_grid, figsize=(n_cols_grid * 5.5, n_rows_grid * 4.5))
    fig.patch.set_facecolor('#1E1E1E')
    
    # AQUÍ ESTÁ EL CAMBIO: Título ajustado para estar más cerca
    fig.suptitle('Análisis de Distribución por Clase', fontsize=22, color='white', weight='bold', y=0.99)

    axes = axes.flatten()

    for i, col in enumerate(df_num.columns):
        ax = axes[i]
        ax.set_facecolor('#1E1E1E')

        # AQUÍ ESTÁ EL CAMBIO: 'hue' añadido y 'legend=False' para eliminar la advertencia
        sns.violinplot(
            x=categoria_col, y=col, data=df, ax=ax,
            palette=palette_ajustada, inner='box', linewidth=1.5,
            hue=categoria_col, legend=False
        )

        ax.set_title(col.replace('_', ' ').capitalize(), fontsize=14, color='white', weight='bold', pad=12)
        ax.set_xlabel('')
        ax.set_ylabel('')
        ax.tick_params(axis='x', colors='white', labelrotation=0)
        ax.tick_params(axis='y', colors='white')
        
        ax.spines['top'].set_visible(False)
        ax.spines['right'].set_visible(False)
        ax.spines['left'].set_color('#444444')
        ax.spines['bottom'].set_color('#444444')
        ax.grid(axis='y', linestyle='--', alpha=0.2)

    for j in range(i + 1, len(axes)):
        axes[j].set_visible(False)

    plt.tight_layout(rect=[0, 0, 1, 0.95])
    plt.show()


# --- CÓMO USARLA ---
visualizar_violines(df)

variables_trans = df.columns.to_list()
variables_trans.remove("Class")

transformer = PowerTransformer(method="yeo-johnson", standardize=False)

transformer.fit(df[variables_trans])

PowerTransformer(standardize=False)

df[variables_trans] = transformer.transform(df[variables_trans])

n = len(variables_trans)

ncols = 8
nrows = math.ceil(n / ncols)

# ---------- FIGURA 1: HISTOGRAMAS ----------
fig1, axes1 = plt.subplots(nrows, ncols, figsize=(ncols*2.0, nrows*1.8))  # plots más compactos
axes1 = axes1.flatten() if nrows*ncols > 1 else [axes1]

for i, col in enumerate(variables_trans):
    ax = axes1[i]
    data = df[col].dropna()
    ax.hist(data, bins=20, edgecolor='white')
    ax.set_title(col, fontsize=9)
    ax.tick_params(labelsize=8)

# apaga ejes sobrantes si hay
for j in range(i+1, nrows*ncols):
    axes1[j].axis('off')

fig1.suptitle('Histogramas', fontsize=12, y=1.02)
fig1.tight_layout()
plt.show()

# ---------- FIGURA 2: Q-Q PLOTS ----------
fig2, axes2 = plt.subplots(nrows, ncols, figsize=(ncols*2.0, nrows*1.8))
axes2 = axes2.flatten() if nrows*ncols > 1 else [axes2]

for i, col in enumerate(variables_trans):
    ax = axes2[i]
    data = df[col].dropna()
    stats.probplot(data, dist="norm", plot=ax)
    ax.set_title(col, fontsize=9)
    ax.tick_params(labelsize=8)

for j in range(i+1, nrows*ncols):
    axes2[j].axis('off')

fig2.suptitle('Q-Q plots vs Normal', fontsize=12, y=1.02)
fig2.tight_layout()
plt.show()

for column in df.select_dtypes(include=[ 'object', 'bool']).columns:
  display(column, pd.crosstab(index=df[column], columns='% observations', normalize='columns') * 100)

'Class'

label_encoder = LabelEncoder()
df["Class"] = label_encoder.fit_transform(df['Class'])
df

df.corr()

data_df_numeric = df.select_dtypes(include=[np.number])

data_df_corr = data_df_numeric.corr()

plt.figure(figsize=(18,12))
sns.heatmap(data_df_corr, annot=True, linewidths=0.5)
plt.show()

if 'Class' in df.columns:
    numeric_cols = df.select_dtypes(include=[float, int]).columns
    numeric_cols = numeric_cols.drop('Class', errors='ignore')
    for col in numeric_cols[:5]:
        plt.figure(figsize=(6, 3))
        sns.boxplot(x='Class', y=col, data=df)
        plt.title(f"Boxplot de {col} por clase")
        plt.tight_layout()
        plt.show()
else:
    print("La columna 'Class' no existe en el DataFrame.")

for col in numeric_cols[:5]:
    plt.figure(figsize=(5,3))
    sns.kdeplot(data=df, x=col, hue="Class", fill=True)
    plt.title(f"Densidad de {col}")
    plt.tight_layout()
    plt.show()

z_scores = np.abs(stats.zscore(df[numeric_cols], nan_policy='omit'))
outlier_mask = (z_scores > 3)
outlier_counts = outlier_mask.sum(axis=0)
outlier_counts_series = pd.Series(outlier_counts, index=numeric_cols)
outlier_counts_series.sort_values(ascending=False).head(10)

_MFCC_Mean_6                                      8
_MFCC_Mean_8                                      6
_MFCC_Mean_10                                     5
_Lowenergy_Mean                                   4
_MFCC_Mean_2                                      4
_HarmonicChangeDetectionFunction_PeriodEntropy    4
_MFCC_Mean_4                                      3
_MFCC_Mean_5                                      3
_MFCC_Mean_3                                      3
_MFCC_Mean_11                                     3
dtype: int64

Q1 = df[numeric_cols].quantile(0.25)
Q3 = df[numeric_cols].quantile(0.75)
IQR = Q3 - Q1

is_outlier = ((df[numeric_cols] < (Q1 - 1.5 * IQR)) |
              (df[numeric_cols] > (Q3 + 1.5 * IQR)))
# cuenta
is_outlier.sum().sort_values(ascending=False).head(10)

_MFCC_Mean_6                                      17
_Fluctuation_Mean                                 16
_HarmonicChangeDetectionFunction_PeriodEntropy    15
_MFCC_Mean_13                                     14
_MFCC_Mean_10                                     13
_MFCC_Mean_1                                      13
_AttackTime_Mean                                  13
_MFCC_Mean_8                                      12
_MFCC_Mean_9                                      11
_MFCC_Mean_5                                      11
dtype: int64

print(df["Class"].unique())

[2 1 3 0]

dup_count = df.duplicated().sum()
print("Duplicados:", dup_count)
if dup_count > 0:
    df = df.drop_duplicates()

Duplicados: 12

stds = df[numeric_cols].std()
zero_std = stds[stds == 0].index.tolist()
print("Columnas constantes:", zero_std)
df = df.drop(columns=zero_std)

Columnas constantes: []

df_original = df.to_csv("turkish_music_emotion_raw.csv", index=False)
df.to_csv("turkish_music_emotion_cleaned.csv", index=False)

X = df.loc[:,df.columns != "Class"]
Y = df.loc[:,df.columns == "Class"]
X

X_train, X_test, Y_train, Y_test = train_test_split(X,Y, test_size=0.3,random_state=42)

standardscaler = StandardScaler()
x_train = standardscaler.fit_transform(X_train)
x_test = standardscaler.transform(X_test)

logrregression = LogisticRegression(random_state=0)
logrregression.fit(x_train,Y_train.values.ravel())
y_pred = logrregression.predict(x_test)
print(y_pred)
print(Y_test)

[3 2 1 2 0 1 1 3 1 1 0 2 1 3 3 1 2 3 3 3 0 0 2 1 0 3 0 0 2 0 3 2 3 0 0 2 3
 0 3 1 1 1 2 1 2 3 3 0 3 3 0 2 3 2 3 1 3 2 2 2 1 3 0 3 0 2 2 0 0 1 0 2 1 2
 2 2 1 0 2 3 1 1 0 3 1 3 3 1 3 0 2 0 2 3 0 2 2 0 2 2 3 0 0 2 1 3 0 0 2 3 0
 1 3 1 1 3 0]
     Class
279      3
46       2
172      1
42       2
359      0
..     ...
252      3
216      3
113      1
17       2
301      0

[117 rows x 1 columns]

accuracy = accuracy_score(Y_test,y_pred)
print(accuracy)

0.7777777777777778

cm = confusion_matrix(Y_test,y_pred)
cm

array([[26,  0,  1,  2],
       [ 2, 23,  1,  2],
       [ 1,  0, 25, 11],
       [ 1,  2,  3, 17]])

model = KNeighborsClassifier()
model.fit(x_train, Y_train.values.ravel())
y_pred = model.predict(x_test)
print(accuracy_score(Y_test,y_pred))

0.6495726495726496

cm = confusion_matrix(Y_test,y_pred)
cm

array([[26,  2,  0,  1],
       [ 1, 27,  0,  0],
       [ 7,  4, 15, 11],
       [ 6,  4,  5,  8]])

model = KNeighborsClassifier(metric="manhattan")
model.fit(x_train, Y_train.values.ravel())
y_pred = model.predict(x_test)
print(accuracy_score(Y_test,y_pred))

0.6581196581196581

cm = confusion_matrix(Y_test,y_pred)
cm

array([[26,  1,  0,  2],
       [ 1, 27,  0,  0],
       [ 6,  2, 14, 15],
       [ 6,  3,  4, 10]])

svc = SVC(kernel='linear')
svc.fit(x_train,Y_train.values.ravel())
y_pred = svc.predict(x_test)
print(accuracy_score(Y_test,y_pred))

0.7264957264957265

cm = confusion_matrix(Y_test,y_pred)
cm

array([[28,  0,  1,  0],
       [ 0, 24,  1,  3],
       [ 4,  0, 20, 13],
       [ 3,  3,  4, 13]])

svc = SVC(kernel='rbf')
svc.fit(x_train,Y_train.values.ravel())
y_pred = svc.predict(x_test)
print(accuracy_score(Y_test,y_pred))

0.7521367521367521

cm = confusion_matrix(Y_test,y_pred)
cm

array([[23,  1,  0,  5],
       [ 0, 25,  1,  2],
       [ 0,  0, 25, 12],
       [ 1,  2,  5, 15]])

gnb = GaussianNB()
gnb.fit(x_train,Y_train.values.ravel())
y_pred = gnb.predict(x_test)
print(accuracy_score(Y_test,y_pred))

0.7863247863247863

cm = confusion_matrix(Y_test,y_pred)
cm

array([[23,  2,  1,  3],
       [ 0, 25,  0,  3],
       [ 0,  1, 29,  7],
       [ 0,  3,  5, 15]])

dtc = DecisionTreeClassifier(criterion="entropy")
dtc.fit(x_train,Y_train.values.ravel())
y_pred = dtc.predict(x_test)
print(accuracy_score(Y_test,y_pred))

0.6752136752136753

cm = confusion_matrix(Y_test,y_pred)
cm

array([[24,  1,  0,  4],
       [ 2, 22,  2,  2],
       [ 2,  3, 21, 11],
       [ 3,  4,  4, 12]])

rfc = RandomForestClassifier()
rfc.fit(x_train,Y_train.values.ravel())
y_pred = rfc.predict(x_test)
print(accuracy_score(Y_test,y_pred))

0.8205128205128205

cm = confusion_matrix(Y_test,y_pred)
cm

array([[26,  0,  0,  3],
       [ 1, 26,  0,  1],
       [ 0,  1, 26, 10],
       [ 1,  2,  2, 18]])

pca = PCA(n_components=8)
pca.fit(X)
x_pca = pca.transform(X)
transformed = pd.DataFrame(x_pca)
X=transformed
X

X_train, X_test, Y_train, Y_test = train_test_split(X,Y, test_size=0.3,random_state=42)

x_train = standardscaler.fit_transform(X_train)
x_test = standardscaler.transform(X_test)

classifier = KNeighborsClassifier()
classifier.fit(x_train, Y_train.values.ravel())

KNeighborsClassifier()

y_pred = classifier.predict(x_test)
print(accuracy_score(Y_test,y_pred))

0.48717948717948717

scaler = StandardScaler()
X_train_scaled = scaler.fit_transform(X_train)
X_test_scaled = scaler.transform(X_test)


pipeline = Pipeline([
    ('scaler', StandardScaler()),
    ('classifier', LogisticRegression(solver='lbfgs', max_iter=1000))
])

param_grid = {'classifier__C': [0.001, 0.01, 0.1, 0.15, 0.2, 0.25, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1, 10, 100]}

grid_search = GridSearchCV(pipeline, param_grid, cv=10, scoring='accuracy')
grid_search.fit(X_train_scaled, Y_train)

# Get the best hyperparameters and accuracy
best_C = grid_search.best_params_['classifier__C']
best_accuracy = grid_search.best_score_
print("Best C:", best_C)
print("Best Accuracy:", best_accuracy)

Categoría	Herramientas
💻 Lenguaje
📊 Data Analysis
🤖 ML Framework
🔄 Versioning
📈 Visualization

Emoción	Características Musicales	Aplicaciones
😊 Happy	Tempo rápido, tonalidad mayor	Playlists energéticas, marketing
😢 Sad	Tempo lento, tonalidad menor	Terapia musical, cine
😠 Angry	Alta intensidad, disonancia	Detección de contenido, videojuegos
😌 Relax	Tempo moderado, armonías suaves	Meditación, ambientes

#	Componente	Estado	Peso
1️⃣	ML Canvas & Propuesta de Valor	🔄 En Progreso	15%
2️⃣	EDA & Data Cleaning	🔄 En Progreso	20%
3️⃣	Feature Engineering	🔄 En Progreso	15%
4️⃣	Data Versioning (DVC)	🔄 En Progreso	15%
5️⃣	Model Development	🔄 En Progreso	20%
6️⃣	Evaluation & Documentation	🔄 En Progreso	15%

	Tipo de Dato	Valores No Nulos	% Nulos	Valores Únicos	Cardinalidad (%)
Class	object	400	0.00%	4	1.00%
_Chromagram_Mean_6	float64	400	0.00%	241	60.25%
_Spectralspread_Mean	float64	400	0.00%	388	97.00%
_Spectralskewness_Mean	float64	400	0.00%	357	89.25%
_Spectralkurtosis_Mean	float64	400	0.00%	381	95.25%
_Spectralflatness_Mean	float64	400	0.00%	94	23.50%
_EntropyofSpectrum_Mean	float64	400	0.00%	134	33.50%
_Chromagram_Mean_1	float64	400	0.00%	259	64.75%
_Chromagram_Mean_2	float64	400	0.00%	240	60.00%
_Chromagram_Mean_3	float64	400	0.00%	263	65.75%
_Chromagram_Mean_4	float64	400	0.00%	240	60.00%
_Chromagram_Mean_5	float64	400	0.00%	286	71.50%
_Chromagram_Mean_7	float64	400	0.00%	243	60.75%
_Brightness_Mean	float64	400	0.00%	277	69.25%
_Chromagram_Mean_8	float64	400	0.00%	269	67.25%
_Chromagram_Mean_9	float64	400	0.00%	262	65.50%
_Chromagram_Mean_10	float64	400	0.00%	245	61.25%
_Chromagram_Mean_11	float64	400	0.00%	273	68.25%
_Chromagram_Mean_12	float64	400	0.00%	252	63.00%
_HarmonicChangeDetectionFunction_Mean	float64	400	0.00%	178	44.50%
_HarmonicChangeDetectionFunction_Std	float64	400	0.00%	159	39.75%
_HarmonicChangeDetectionFunction_Slope	float64	400	0.00%	237	59.25%
_HarmonicChangeDetectionFunction_PeriodFreq	float64	400	0.00%	40	10.00%
_HarmonicChangeDetectionFunction_PeriodAmp	float64	400	0.00%	196	49.00%
_Spectralcentroid_Mean	float64	400	0.00%	388	97.00%
_Pulseclarity_Mean	float64	400	0.00%	266	66.50%
_RMSenergy_Mean	float64	400	0.00%	196	49.00%
_MFCC_Mean_8	float64	400	0.00%	273	68.25%
_Lowenergy_Mean	float64	400	0.00%	166	41.50%
_Fluctuation_Mean	float64	400	0.00%	377	94.25%
_Tempo_Mean	float64	400	0.00%	388	97.00%
_MFCC_Mean_1	float64	400	0.00%	354	88.50%
_MFCC_Mean_2	float64	400	0.00%	347	86.75%
_MFCC_Mean_3	float64	400	0.00%	319	79.75%
_MFCC_Mean_4	float64	400	0.00%	316	79.00%
_MFCC_Mean_5	float64	400	0.00%	297	74.25%
_MFCC_Mean_6	float64	400	0.00%	297	74.25%
_MFCC_Mean_7	float64	400	0.00%	304	76.00%
_MFCC_Mean_9	float64	400	0.00%	278	69.50%
_Eventdensity_Mean	float64	400	0.00%	163	40.75%
_MFCC_Mean_10	float64	400	0.00%	271	67.75%
_MFCC_Mean_11	float64	400	0.00%	253	63.25%
_MFCC_Mean_12	float64	400	0.00%	272	68.00%
_MFCC_Mean_13	float64	400	0.00%	259	64.75%
_Roughness_Mean	float64	400	0.00%	388	97.00%
_Roughness_Slope	float64	400	0.00%	292	73.00%
_Zero-crossingrate_Mean	float64	400	0.00%	388	97.00%
_AttackTime_Mean	float64	400	0.00%	61	15.25%
_AttackTime_Slope	float64	400	0.00%	274	68.50%
_Rolloff_Mean	float64	400	0.00%	388	97.00%
_HarmonicChangeDetectionFunction_PeriodEntropy	float64	400	0.00%	26	6.50%

	_RMSenergy_Mean	_Lowenergy_Mean	_Fluctuation_Mean	_Tempo_Mean	_MFCC_Mean_1	_MFCC_Mean_2	_MFCC_Mean_3	_MFCC_Mean_4	_MFCC_Mean_5	_MFCC_Mean_6	...	_Chromagram_Mean_9	_Chromagram_Mean_10	_Chromagram_Mean_11	_Chromagram_Mean_12	_HarmonicChangeDetectionFunction_Mean	_HarmonicChangeDetectionFunction_Std	_HarmonicChangeDetectionFunction_Slope	_HarmonicChangeDetectionFunction_PeriodFreq	_HarmonicChangeDetectionFunction_PeriodAmp	_HarmonicChangeDetectionFunction_PeriodEntropy
count	400.000000	400.000000	400.000000	400.000000	400.000000	400.000000	400.000000	400.000000	400.000000	400.000000	...	400.000000	400.000000	400.000000	400.000000	400.000000	400.000000	400.000000	400.000000	400.000000	400.000000
mean	0.134650	0.553605	7.145932	123.682020	2.456422	0.071890	0.488065	0.030465	0.178897	0.038307	...	0.354632	0.590975	0.342340	0.385620	0.328213	0.192997	-0.000157	1.762288	0.769690	0.966712
std	0.064368	0.050750	2.280145	34.234344	0.799262	0.537865	0.294607	0.275839	0.195230	0.203754	...	0.334976	0.357981	0.315808	0.348117	0.055520	0.047092	0.104743	0.930352	0.072107	0.003841
min	0.010000	0.302000	3.580000	48.284000	0.323000	-3.484000	-0.870000	-1.636000	-0.494000	-0.916000	...	0.000000	0.000000	0.000000	0.000000	0.112000	0.060000	-0.285000	0.187000	0.530000	0.939000
25%	0.085000	0.523000	5.859500	101.490250	1.948500	-0.262750	0.281250	-0.117000	0.061250	-0.078250	...	0.066750	0.264500	0.059500	0.060750	0.290750	0.160000	-0.058000	0.961000	0.725000	0.965000
50%	0.128000	0.553000	6.734000	120.132500	2.389500	0.068500	0.464500	0.044500	0.181000	0.049500	...	0.247000	0.612000	0.247000	0.296500	0.333000	0.190000	-0.002000	1.682000	0.786000	0.967000
75%	0.174000	0.583250	7.823500	148.986250	2.860250	0.413250	0.686000	0.198250	0.288500	0.151250	...	0.612000	1.000000	0.565250	0.670750	0.367250	0.226000	0.063250	2.243000	0.824000	0.969000
max	0.431000	0.703000	23.475000	195.026000	5.996000	1.937000	1.622000	1.126000	1.055000	0.799000	...	1.000000	1.000000	1.000000	1.000000	0.488000	0.340000	0.442000	4.486000	0.908000	0.977000

🎵 Operaciones de Aprendizaje Automático (MLOps)¶

👨‍🏫 Equipo Docente¶

🎯 Dataset del Proyecto¶

🎵 Turkish Music Emotion Dataset¶

👥 Equipo de Desarrollo¶

David Cruz Beltrán

Javier Augusto Rebull Saucedo

Sandra Luz Cervantes Espinoza

📋 Objetivos del Proyecto¶

🎯 Análisis¶

🛠️ Desarrollo¶

🤖 Modelado¶

🔧 Stack Tecnológico¶

🎼 Clasificación de Emociones¶

🚀 Metodología MLOps¶

📈 Entregables¶

🔗 Enlaces del Proyecto¶

🎵 Transformando emociones musicales en conocimiento mediante MLOps 🤖¶

📊 Proceso de Análisis de Datos¶

🔍 Análisis Exploratorio (EDA)¶

🧹 Preprocesamiento¶

⚙️ Ingeniería de Características¶

🚀 Entrenamiento y Evaluación del Modelo¶

🔬 1. Análisis Exploratorio de Datos (EDA)¶

📊 Estadísticas Descriptivas¶

Ingeniería de características¶

Detección de valores atípicos (outliers)¶

Limpieza y tratamiento de datos¶

	count	mean	median	std	min	25%	50%	75%	max	skew	kurtosis
_RMSenergy_Mean	400.00	0.13	0.13	0.06	0.01	0.09	0.13	0.17	0.43	0.71	0.62
_Lowenergy_Mean	400.00	0.55	0.55	0.05	0.30	0.52	0.55	0.58	0.70	-0.39	2.17
_Fluctuation_Mean	400.00	7.15	6.73	2.28	3.58	5.86	6.73	7.82	23.48	2.89	12.73
_Tempo_Mean	400.00	123.68	120.13	34.23	48.28	101.49	120.13	148.99	195.03	0.12	-0.63
_MFCC_Mean_1	400.00	2.46	2.39	0.80	0.32	1.95	2.39	2.86	6.00	0.87	1.95
_MFCC_Mean_2	400.00	0.07	0.07	0.54	-3.48	-0.26	0.07	0.41	1.94	-0.68	4.91
_MFCC_Mean_3	400.00	0.49	0.46	0.29	-0.87	0.28	0.46	0.69	1.62	0.11	1.19
_MFCC_Mean_4	400.00	0.03	0.04	0.28	-1.64	-0.12	0.04	0.20	1.13	-0.70	4.08
_MFCC_Mean_5	400.00	0.18	0.18	0.20	-0.49	0.06	0.18	0.29	1.05	0.08	1.19
_MFCC_Mean_6	400.00	0.04	0.05	0.20	-0.92	-0.08	0.05	0.15	0.80	-0.28	2.58
_MFCC_Mean_7	400.00	0.06	0.07	0.18	-0.94	-0.04	0.07	0.17	0.57	-0.79	3.35
_MFCC_Mean_8	400.00	0.04	0.04	0.17	-0.74	-0.05	0.04	0.13	0.73	-0.11	2.86
_MFCC_Mean_9	400.00	0.02	0.02	0.16	-0.62	-0.07	0.02	0.12	0.54	-0.13	1.29
_MFCC_Mean_10	400.00	0.03	0.03	0.15	-0.54	-0.06	0.03	0.13	0.51	-0.25	1.08
_MFCC_Mean_11	400.00	0.03	0.04	0.14	-0.49	-0.04	0.04	0.11	0.49	-0.30	0.91
_MFCC_Mean_12	400.00	0.02	0.02	0.13	-0.42	-0.06	0.02	0.09	0.35	-0.35	0.40
_MFCC_Mean_13	400.00	0.02	0.04	0.13	-0.62	-0.05	0.04	0.10	0.54	-0.49	2.02
_Roughness_Mean	400.00	527.68	367.58	521.22	0.94	169.19	367.58	734.37	3899.85	2.08	6.73
_Roughness_Slope	400.00	0.07	0.07	0.17	-0.53	-0.03	0.07	0.17	0.58	-0.02	0.42
_Zero-crossingrate_Mean	400.00	997.25	893.49	524.90	149.49	592.27	893.49	1303.49	3147.91	0.93	0.69
_AttackTime_Mean	400.00	0.03	0.03	0.02	0.01	0.02	0.03	0.03	0.17	3.38	17.11
_AttackTime_Slope	400.00	-0.00	0.01	0.15	-0.47	-0.09	0.01	0.09	0.60	-0.01	0.47
_Rolloff_Mean	400.00	5691.07	5648.63	2293.40	887.15	3933.55	5648.63	7355.89	11508.30	0.10	-0.61
_Eventdensity_Mean	400.00	2.78	2.77	1.33	0.23	1.74	2.77	3.69	7.95	0.48	0.13
_Pulseclarity_Mean	400.00	0.25	0.22	0.16	0.01	0.13	0.22	0.33	0.86	1.16	1.21
_Brightness_Mean	400.00	0.43	0.45	0.13	0.05	0.35	0.45	0.53	0.74	-0.40	-0.07
_Spectralcentroid_Mean	400.00	2581.17	2547.68	863.52	606.52	1981.56	2547.68	3182.57	5326.38	0.23	-0.13
_Spectralspread_Mean	400.00	3082.39	3150.95	767.65	814.82	2506.77	3150.95	3684.33	4721.48	-0.26	-0.43
_Spectralskewness_Mean	400.00	1.87	1.69	0.88	0.39	1.33	1.69	2.18	7.86	2.28	9.12
_Spectralkurtosis_Mean	400.00	7.35	5.22	8.62	1.93	3.88	5.22	7.85	122.00	7.97	89.68
_Spectralflatness_Mean	400.00	0.05	0.05	0.03	0.01	0.03	0.05	0.06	0.21	1.49	5.17
_EntropyofSpectrum_Mean	400.00	0.87	0.88	0.04	0.74	0.85	0.88	0.90	0.94	-0.98	0.96
_Chromagram_Mean_1	400.00	0.35	0.27	0.32	0.00	0.06	0.27	0.55	1.00	0.70	-0.72
_Chromagram_Mean_2	400.00	0.25	0.14	0.29	0.00	0.02	0.14	0.40	1.00	1.21	0.43
_Chromagram_Mean_3	400.00	0.37	0.29	0.32	0.00	0.08	0.29	0.58	1.00	0.69	-0.73
_Chromagram_Mean_4	400.00	0.21	0.10	0.25	0.00	0.02	0.10	0.32	1.00	1.49	1.48
_Chromagram_Mean_5	400.00	0.35	0.27	0.30	0.00	0.09	0.27	0.54	1.00	0.78	-0.46
_Chromagram_Mean_6	400.00	0.26	0.14	0.29	0.00	0.02	0.14	0.45	1.00	1.11	0.19
_Chromagram_Mean_7	400.00	0.24	0.14	0.28	0.00	0.03	0.14	0.36	1.00	1.33	0.93
_Chromagram_Mean_8	400.00	0.39	0.30	0.33	0.00	0.10	0.30	0.64	1.00	0.57	-0.96
_Chromagram_Mean_9	400.00	0.35	0.25	0.33	0.00	0.07	0.25	0.61	1.00	0.76	-0.77
_Chromagram_Mean_10	400.00	0.59	0.61	0.36	0.00	0.26	0.61	1.00	1.00	-0.22	-1.44
_Chromagram_Mean_11	400.00	0.34	0.25	0.32	0.00	0.06	0.25	0.57	1.00	0.71	-0.69
_Chromagram_Mean_12	400.00	0.39	0.30	0.35	0.00	0.06	0.30	0.67	1.00	0.55	-1.12
_HarmonicChangeDetectionFunction_Mean	400.00	0.33	0.33	0.06	0.11	0.29	0.33	0.37	0.49	-0.47	0.49
_HarmonicChangeDetectionFunction_Std	400.00	0.19	0.19	0.05	0.06	0.16	0.19	0.23	0.34	0.25	-0.11
_HarmonicChangeDetectionFunction_Slope	400.00	-0.00	-0.00	0.10	-0.28	-0.06	-0.00	0.06	0.44	0.20	0.93
_HarmonicChangeDetectionFunction_PeriodFreq	400.00	1.76	1.68	0.93	0.19	0.96	1.68	2.24	4.49	0.40	-0.29
_HarmonicChangeDetectionFunction_PeriodAmp	400.00	0.77	0.79	0.07	0.53	0.72	0.79	0.82	0.91	-0.76	0.18
_HarmonicChangeDetectionFunction_PeriodEntropy	400.00	0.97	0.97	0.00	0.94	0.96	0.97	0.97	0.98	-1.48	7.57

	Class	_RMSenergy_Mean	_Lowenergy_Mean	_Fluctuation_Mean	_Tempo_Mean	_MFCC_Mean_1	_MFCC_Mean_2	_MFCC_Mean_3	_MFCC_Mean_4	_MFCC_Mean_5	...	_Chromagram_Mean_9	_Chromagram_Mean_10	_Chromagram_Mean_11	_Chromagram_Mean_12	_HarmonicChangeDetectionFunction_Mean	_HarmonicChangeDetectionFunction_Std	_HarmonicChangeDetectionFunction_Slope	_HarmonicChangeDetectionFunction_PeriodFreq	_HarmonicChangeDetectionFunction_PeriodAmp	_HarmonicChangeDetectionFunction_PeriodEntropy
0	2	0.046255	1.147712	0.785693	49.459481	1.883949	0.377859	0.865352	0.079494	0.219944	...	0.269816	1.136109	0.007925	0.091497	0.537622	0.200511	0.017918	0.807322	6.762849	5.938439e+30
1	2	0.095623	0.727046	0.764910	52.943619	1.900475	0.545123	0.767637	0.433917	0.549911	...	0.001995	1.136109	-0.000000	0.487793	0.460935	0.169693	-0.083691	1.931502	12.210196	5.010738e+30
2	2	0.041457	1.302833	0.793466	65.444342	1.512479	0.986259	0.494373	0.354595	0.285253	...	0.147740	0.824692	0.015702	0.491675	0.821232	0.222127	0.129709	1.179722	11.586369	3.993609e+30
3	2	0.101274	1.185439	0.792824	29.467297	1.534865	1.775386	0.600987	0.380042	0.010997	...	0.036190	1.136109	0.134976	0.424863	0.851187	0.202856	0.041560	0.319833	15.087478	5.302789e+30
4	2	0.056968	1.147712	0.789386	37.016335	1.656768	0.234033	0.795458	0.098256	0.430168	...	0.003979	0.428451	0.447120	0.000999	0.615226	0.200511	0.087067	0.617198	10.533691	2.839124e+30
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
395	0	0.121233	1.107620	0.744632	58.198474	1.582646	0.065509	0.703238	0.046522	0.263501	...	0.248283	0.935896	0.275238	0.110760	0.555823	0.120580	0.116536	1.658183	27.746435	5.611702e+30
396	0	0.122175	0.878151	0.740478	63.229758	1.517594	-0.145470	0.338298	-0.010970	0.028981	...	0.019488	1.136109	0.359050	0.009898	0.345856	0.110819	0.140000	1.931502	29.365195	5.010738e+30
397	0	0.127222	1.044541	0.733938	50.607703	1.081312	0.600708	0.858658	-0.109992	0.242721	...	0.048665	0.189290	0.213246	0.091497	0.423771	0.132968	0.108024	1.931502	22.030962	3.773116e+30
398	0	0.104013	1.092413	0.728113	44.628036	1.221377	-0.205021	0.680127	0.090942	0.205078	...	0.115906	1.136109	0.222443	0.122376	0.442152	0.123533	0.060080	1.931502	21.187036	5.611702e+30
399	0	0.071178	0.817475	0.746010	55.607331	1.318279	-0.013976	0.814626	-0.020891	0.342518	...	0.087529	1.136109	0.084520	0.031917	0.271693	0.097658	0.006988	0.541146	25.344224	4.473616e+30

	Class	_RMSenergy_Mean	_Lowenergy_Mean	_Fluctuation_Mean	_Tempo_Mean	_MFCC_Mean_1	_MFCC_Mean_2	_MFCC_Mean_3	_MFCC_Mean_4	_MFCC_Mean_5	...	_Chromagram_Mean_9	_Chromagram_Mean_10	_Chromagram_Mean_11	_Chromagram_Mean_12	_HarmonicChangeDetectionFunction_Mean	_HarmonicChangeDetectionFunction_Std	_HarmonicChangeDetectionFunction_Slope	_HarmonicChangeDetectionFunction_PeriodFreq	_HarmonicChangeDetectionFunction_PeriodAmp	_HarmonicChangeDetectionFunction_PeriodEntropy
Class	1.000000	-0.300955	0.174327	0.392882	-0.056205	0.362134	0.068330	-0.118221	-0.050932	-0.045488	...	0.028964	-0.111679	-0.054256	0.044204	0.314104	0.669975	0.019715	-0.201242	-0.597808	-0.063368
_RMSenergy_Mean	-0.300955	1.000000	-0.281635	-0.177114	-0.014712	-0.171867	-0.025694	0.075546	0.004651	-0.049314	...	0.128949	0.131721	0.163869	0.029282	-0.014093	-0.369574	-0.091747	0.120925	0.355068	-0.032966
_Lowenergy_Mean	0.174327	-0.281635	1.000000	0.155663	-0.042172	0.118139	0.133686	-0.067116	0.056634	0.018319	...	0.074131	-0.025039	0.066540	-0.034081	0.023246	0.207735	0.223859	-0.088219	-0.196982	-0.021062
_Fluctuation_Mean	0.392882	-0.177114	0.155663	1.000000	-0.106790	0.124968	0.147555	-0.130878	0.120289	-0.055065	...	-0.028560	-0.083241	0.001803	0.060133	0.338422	0.452092	0.104434	-0.069151	-0.306285	-0.040611
_Tempo_Mean	-0.056205	-0.014712	-0.042172	-0.106790	1.000000	-0.075108	0.083978	0.017875	0.025673	0.072101	...	0.060861	0.059757	0.035827	0.015775	-0.089613	-0.137069	-0.058962	0.032283	0.059248	0.116209
_MFCC_Mean_1	0.362134	-0.171867	0.118139	0.124968	-0.075108	1.000000	0.040707	0.067129	0.049199	-0.107084	...	-0.177787	-0.094558	-0.177474	-0.060446	-0.004390	0.398240	0.020684	-0.126185	-0.431880	-0.072597
_MFCC_Mean_2	0.068330	-0.025694	0.133686	0.147555	0.083978	0.040707	1.000000	0.047537	0.358970	0.177046	...	0.003020	0.009304	-0.117024	-0.025309	-0.041602	0.136278	0.147273	-0.052555	-0.165532	0.033975
_MFCC_Mean_3	-0.118221	0.075546	-0.067116	-0.130878	0.017875	0.067129	0.047537	1.000000	0.190250	0.104222	...	-0.079648	0.009654	-0.058884	-0.023069	-0.157477	-0.079437	0.076963	-0.040040	-0.026490	-0.058232
_MFCC_Mean_4	-0.050932	0.004651	0.056634	0.120289	0.025673	0.049199	0.358970	0.190250	1.000000	0.272867	...	-0.047560	0.062361	-0.090089	-0.025413	-0.075269	-0.004211	0.127470	-0.075101	-0.008205	-0.092063
_MFCC_Mean_5	-0.045488	-0.049314	0.018319	-0.055065	0.072101	-0.107084	0.177046	0.104222	0.272867	1.000000	...	0.057164	0.005268	0.023767	0.030171	-0.085922	-0.104762	0.037293	0.027463	0.062554	0.013751
_MFCC_Mean_6	-0.037669	-0.038272	-0.006840	-0.112516	0.064635	0.063804	0.181029	0.121178	0.337789	0.379467	...	-0.018752	0.068534	-0.040221	-0.080736	-0.119552	-0.078788	-0.011430	-0.051585	0.040636	-0.067341
_MFCC_Mean_7	-0.044928	-0.073890	-0.007664	-0.204872	0.109782	-0.033569	0.057125	0.097992	0.055655	0.105031	...	0.065463	0.078696	-0.025926	-0.092893	-0.142082	-0.126186	0.014191	-0.080202	0.030478	0.041139
_MFCC_Mean_8	0.029137	-0.028576	0.002668	-0.009350	0.084302	0.023452	0.006055	0.048145	-0.048309	-0.006843	...	0.105767	0.045640	-0.016686	-0.089824	-0.054623	-0.065312	-0.082687	-0.004595	0.004550	0.004134
_MFCC_Mean_9	0.032591	-0.061882	0.139184	-0.019740	0.089347	0.018528	0.133188	-0.023552	0.040513	-0.043807	...	0.011716	-0.005635	0.017750	-0.119475	-0.086143	0.017109	0.091577	-0.070629	-0.074413	-0.028158
_MFCC_Mean_10	-0.020621	-0.024011	0.061075	0.092923	0.089506	-0.045888	0.054751	0.006540	0.005815	0.044985	...	0.000884	-0.037011	0.001084	-0.096384	-0.044083	0.041982	0.036067	-0.047205	-0.047720	0.012345
_MFCC_Mean_11	-0.075638	0.028033	-0.062946	-0.045076	0.066297	-0.160400	0.114932	-0.119775	-0.011914	0.098805	...	0.014056	0.030343	0.056202	0.065739	0.003975	-0.022370	-0.059868	-0.009596	0.085333	0.069530
_MFCC_Mean_12	0.028299	-0.040536	0.060556	0.047515	0.022166	0.016254	0.005449	0.003575	-0.034378	-0.072253	...	-0.052725	-0.008822	0.024351	0.098350	0.098797	0.082483	0.013647	-0.019403	-0.011682	0.079298
_MFCC_Mean_13	-0.067013	-0.067257	0.103279	-0.047603	0.068089	0.050632	0.133568	0.031047	-0.002487	0.058814	...	-0.086168	0.039328	0.008362	0.132387	-0.026787	0.036666	-0.010579	0.050582	0.005560	0.100646
_Roughness_Mean	-0.394106	0.922060	-0.354121	-0.247770	0.022865	-0.276713	-0.090417	0.080286	-0.045444	-0.019835	...	0.127504	0.154507	0.183550	0.057615	0.003768	-0.459690	-0.097954	0.120594	0.473365	0.000949
_Roughness_Slope	0.144049	-0.101701	0.035621	0.050208	0.036830	0.084562	0.104972	-0.026281	0.062458	-0.005208	...	-0.047285	0.024319	-0.088668	0.103516	0.109817	0.220078	0.086695	-0.086702	-0.164752	0.029399
_Zero-crossingrate_Mean	-0.074970	0.026139	-0.015567	0.069289	0.004459	-0.682747	-0.288590	-0.310513	-0.263036	-0.112658	...	0.205742	0.040806	0.255343	0.179463	0.235178	-0.174276	-0.073436	0.086946	0.331018	0.064745
_AttackTime_Mean	0.480993	-0.444078	0.459084	0.410047	-0.092952	0.192544	0.053319	-0.122228	-0.092701	-0.158953	...	-0.051039	-0.101806	-0.091793	-0.059036	0.079447	0.500102	0.153832	-0.122243	-0.504521	-0.095227
_AttackTime_Slope	-0.110661	0.080719	-0.040061	-0.041544	-0.160343	0.003024	-0.204739	-0.010913	-0.025485	0.057418	...	0.043832	-0.060792	0.096816	0.071297	0.020882	-0.094326	-0.090617	0.089473	0.112861	0.003476
_Rolloff_Mean	-0.156709	0.164842	0.040105	0.110152	0.073996	-0.620981	0.278365	-0.251943	0.125691	0.008531	...	0.199541	0.136537	0.173243	0.109965	0.100547	-0.173244	-0.002296	0.070895	0.257103	0.038293
_Eventdensity_Mean	-0.484784	0.450370	-0.490128	-0.275265	0.101036	-0.448766	-0.205635	-0.048297	-0.063492	0.014219	...	0.168344	0.100157	0.200335	0.065604	0.052560	-0.630868	-0.107198	0.262463	0.681950	0.128439
_Pulseclarity_Mean	-0.372102	0.189068	0.011201	0.033147	0.117209	-0.309492	-0.066375	-0.173008	-0.042196	-0.046745	...	0.176810	0.057012	0.181230	0.070518	0.030180	-0.392775	0.032244	0.156389	0.467441	0.151976
_Brightness_Mean	-0.224204	0.118511	-0.069372	0.003776	0.045190	-0.896142	-0.188497	-0.167434	-0.135427	-0.035968	...	0.193288	0.080318	0.211502	0.126747	0.127376	-0.279805	-0.044780	0.109403	0.377307	0.051659
_Spectralcentroid_Mean	-0.167885	0.154831	0.008884	0.088264	0.065080	-0.739102	0.141171	-0.287106	0.028881	-0.023646	...	0.208623	0.116180	0.191573	0.125356	0.118459	-0.206763	-0.011408	0.085939	0.304879	0.038277
_Spectralspread_Mean	-0.122575	0.197366	0.054832	0.098471	0.067792	-0.437120	0.373816	-0.221671	0.192448	-0.017684	...	0.173048	0.151266	0.135917	0.102115	0.063392	-0.137951	0.018140	0.041517	0.201810	0.008683
_Spectralskewness_Mean	0.201855	-0.144788	0.012160	-0.083325	-0.090589	0.785463	-0.227533	0.233059	-0.099461	-0.055843	...	-0.221493	-0.108266	-0.161126	-0.090343	-0.099755	0.217282	-0.009631	-0.084267	-0.304221	-0.062772
_Spectralkurtosis_Mean	0.178580	-0.160859	-0.013604	-0.120152	-0.092898	0.702285	-0.303576	0.250561	-0.159243	-0.040663	...	-0.215957	-0.121465	-0.155683	-0.099912	-0.105316	0.180356	-0.011200	-0.066685	-0.272795	-0.054307
_Spectralflatness_Mean	-0.058126	0.063555	0.089920	0.043057	0.048616	-0.304261	0.299231	-0.193467	0.085180	-0.059916	...	0.084293	0.077582	0.068592	0.053399	-0.026120	-0.089562	0.026035	0.044578	0.086967	-0.000055
_EntropyofSpectrum_Mean	-0.219075	0.170018	-0.042593	0.033299	0.057302	-0.774152	-0.081432	-0.295714	-0.096215	-0.043524	...	0.243881	0.105560	0.265868	0.175020	0.173405	-0.307949	-0.052169	0.113522	0.433486	0.044032
_Chromagram_Mean_1	-0.032565	0.035289	-0.002613	0.094381	0.060964	-0.064125	-0.058609	-0.134867	-0.054657	0.084882	...	0.093008	-0.178847	0.353609	0.349807	0.257996	-0.038977	-0.115735	0.051425	0.221190	0.001559
_Chromagram_Mean_2	-0.033956	0.071248	-0.042269	0.011150	0.031261	-0.121695	-0.140485	-0.100036	-0.130305	0.025397	...	0.444471	-0.090557	0.256742	0.316175	0.310582	-0.019188	-0.099637	-0.013200	0.238698	0.060834
_Chromagram_Mean_3	-0.117814	0.096670	-0.035443	-0.010909	0.011636	-0.010620	-0.050730	-0.074067	-0.012424	0.020332	...	-0.073335	0.108335	0.165910	-0.094055	0.070674	-0.084338	-0.111410	0.014367	0.136154	0.037803
_Chromagram_Mean_4	-0.035852	0.092103	-0.004291	0.014159	-0.049349	-0.092652	-0.087213	-0.116014	-0.129973	-0.025885	...	0.409477	-0.195227	0.430896	0.156868	0.257533	-0.059484	-0.090352	-0.013615	0.232225	-0.029288
_Chromagram_Mean_5	0.031225	0.019026	-0.101486	0.038371	0.052184	-0.022024	-0.099362	-0.070167	-0.092000	-0.078961	...	-0.005299	0.103886	-0.125068	0.468105	0.117932	0.022755	-0.031876	0.016427	0.057214	-0.005967
_Chromagram_Mean_6	0.069666	-0.017463	0.059919	0.060916	0.007653	-0.064415	-0.146435	-0.045886	-0.133151	0.014098	...	0.228630	-0.082557	0.397089	-0.091050	0.245944	-0.004566	0.011835	0.074252	0.126667	-0.078459
_Chromagram_Mean_7	0.252600	-0.084324	0.034529	0.207115	-0.022048	-0.015167	-0.041778	-0.086944	-0.090960	-0.028213	...	0.267587	-0.134059	0.047911	0.492224	0.359056	0.174767	-0.033643	-0.033668	-0.002329	-0.018215
_Chromagram_Mean_8	0.136248	0.008533	0.031312	0.169993	0.015751	-0.067079	0.059728	-0.048600	0.035155	-0.019915	...	0.295011	-0.065163	0.278067	-0.002241	0.314400	0.058605	-0.058629	0.077626	0.097986	-0.042300
_Chromagram_Mean_9	0.028964	0.128949	0.074131	-0.028560	0.060861	-0.177787	0.003020	-0.079648	-0.047560	0.057164	...	1.000000	0.062536	0.368510	0.136012	0.285409	-0.104406	-0.051001	0.068966	0.254790	0.048431
_Chromagram_Mean_10	-0.111679	0.131721	-0.025039	-0.083241	0.059757	-0.094558	0.009304	0.009654	0.062361	0.005268	...	0.062536	1.000000	0.064108	-0.046829	-0.120855	-0.144962	0.102380	0.000520	0.077126	0.007735
_Chromagram_Mean_11	-0.054256	0.163869	0.066540	0.001803	0.035827	-0.177474	-0.117024	-0.058884	-0.090089	0.023767	...	0.368510	0.064108	1.000000	0.074993	0.227201	-0.161698	-0.091722	0.008532	0.303412	-0.013759
_Chromagram_Mean_12	0.044204	0.029282	-0.034081	0.060133	0.015775	-0.060446	-0.025309	-0.023069	-0.025413	0.030171	...	0.136012	-0.046829	0.074993	1.000000	0.190464	0.047696	-0.050764	-0.001428	0.090829	-0.022490
_HarmonicChangeDetectionFunction_Mean	0.314104	-0.014093	0.023246	0.338422	-0.089613	-0.004390	-0.041602	-0.157477	-0.075269	-0.085922	...	0.285409	-0.120855	0.227201	0.190464	1.000000	0.427895	-0.138121	0.046846	0.122117	0.075975
_HarmonicChangeDetectionFunction_Std	0.669975	-0.369574	0.207735	0.452092	-0.137069	0.398240	0.136278	-0.079437	-0.004211	-0.104762	...	-0.104406	-0.144962	-0.161698	0.047696	0.427895	1.000000	0.027965	-0.260338	-0.752416	-0.130781
_HarmonicChangeDetectionFunction_Slope	0.019715	-0.091747	0.223859	0.104434	-0.058962	0.020684	0.147273	0.076963	0.127470	0.037293	...	-0.051001	0.102380	-0.091722	-0.050764	-0.138121	0.027965	1.000000	-0.056598	-0.105679	-0.154811
_HarmonicChangeDetectionFunction_PeriodFreq	-0.201242	0.120925	-0.088219	-0.069151	0.032283	-0.126185	-0.052555	-0.040040	-0.075101	0.027463	...	0.068966	0.000520	0.008532	-0.001428	0.046846	-0.260338	-0.056598	1.000000	0.363481	0.077017
_HarmonicChangeDetectionFunction_PeriodAmp	-0.597808	0.355068	-0.196982	-0.306285	0.059248	-0.431880	-0.165532	-0.026490	-0.008205	0.062554	...	0.254790	0.077126	0.303412	0.090829	0.122117	-0.752416	-0.105679	0.363481	1.000000	0.074684
_HarmonicChangeDetectionFunction_PeriodEntropy	-0.063368	-0.032966	-0.021062	-0.040611	0.116209	-0.072597	0.033975	-0.058232	-0.092063	0.013751	...	0.048431	0.007735	-0.013759	-0.022490	0.075975	-0.130781	-0.154811	0.077017	0.074684	1.000000

	0	1	2	3	4	5	6	7
0	9.003031e+29	46774.809364	-926.896269	-439.107779	-37.528514	-7.861490	-7.855499	-0.930623
1	-2.739867e+28	-15396.849296	-2013.061404	-762.800698	-55.460648	2.776469	-4.564301	5.838971
2	-1.044527e+30	-49533.852416	-3042.806624	-370.528999	33.938369	13.407332	-5.777505	16.406478
3	2.646531e+29	24525.715202	-4031.572854	-763.691484	40.897985	5.702465	8.370228	-16.578039
4	-2.199013e+30	-150150.116064	-1882.520731	-598.871764	-56.325918	11.225291	2.046949	-7.212007
...	...	...	...	...	...	...	...	...
383	5.735657e+29	41906.372599	-2568.793272	-679.189570	-31.280710	2.245369	5.644715	13.338374
384	-2.739867e+28	18278.013476	-2670.222907	-310.077079	64.282401	-18.758019	1.193700	16.190012
385	-1.265020e+30	-77168.512531	-545.411112	-79.530364	-8.676784	14.326846	2.173908	5.732842
386	5.735657e+29	42590.101548	-1501.956360	-439.647642	-44.886751	7.903570	4.653567	-1.812959
387	-5.645205e+29	-42199.305661	-1706.781605	-477.377457	-11.777035	5.836857	6.503257	12.737869

	n_neighbors	5
	weights	'uniform'
	algorithm	'auto'
	leaf_size	30
	p	2
	metric	'minkowski'
	metric_params	None
	n_jobs	None

	method	'yeo-johnson'
	standardize	False
	copy	True

col_0	% observations
Class
angry	25.0
happy	25.0
relax	25.0
sad	25.0